- Email: [email protected]
Ruptured abdominal aortic aneurysm: The internist as diagnostician
Ruptured Abdominal Aortic Aneurysm: The Internist As Diagnostician FRANK A. LEDERLE, M.D., CONNIE M. PARENTI, M.D.,
EDMUND
P.
PURPOSE: To define the clinical features and assess the frequency and causes of missed diagnoses of ruptured abdominal aortic aneurysm (AAA) in patients initially presenting to internists. PATIENTS: All identified patients with ruptured AAA presenting to internists during a 7V&year period at a large academic medical center. METHOD: Chart review. RESULTS: We identified 23 patients with a ruptured AAA presenting to internists. Most had abdominal pain and tenderness, back or flank pain, and leukocytosis, whereas anemia and profound hypotension (systolic blood pressure below 90 mm Hg) were uncommon at presentation. In 14 cases (61%), the diagnosis of ruptured AAA was initially missed. Nine patients had an interval of 24 hours or more between presentation to the internist and surgery or death. The diagnosis was not made until after shock developed in nine patients who were hemodynamically stable at presentation. Of 17 patients who underwent surgery, 7 of 8 with preoperative shock died, compared with 2 deaths in 9 patients (p c .02) without shock. All six patients who did not have surgery died, yielding an overall mortality of 65% for the series. Ruptured AAAs were most frequently misdiagnosed as urinary tract obstruction or infection, spinal disease, and diverticulitis. Chart review revealed a general lack of physician awareness of the syndromes of contained rupture of AAA and symptomatic unruptured AAA. CONCLUSIONS: In patients with ruptured AAA who present to internists, the diagno-
From the Department of Medicine (FAL, CMP), the Department of Surgery (EPC), Minneapolis Veterans Affairs Medical Center, University of Minnesota, Minneapolis, Minnesota. Requests for reprints should be addressed to Frank A. Lederle, M.D., Minneapolis Veterans Affairs Medical Center, Department of Medicine (III-O), OneVeterans Drive, Minneapolis, Minnesota 55417. Manuscript submitted December 29, 1992, and accepted in revised form March 31, 1993.
February
CHUTE,
M.D.,
Minneapo/is,
Minnesota
sis is often delayed or missed and this appears to adversely effect survival. Internists should familiarize themselves with the presentation and management of ruptured AAA.
T
here are few conditions encountered in the practice of medicine that require accurate diagnosis more urgently than does rupture of an abdominal aortic aneurysm L4AA). In many patients, ruptured AA4 presents as sudden vascular collapse, in which case the need for emergency surgery may be obvious but the likelihood of a successful outcome is low [ll. In other patients, an AAA may become symptomatic prior to rupture or an aneurysmal leak may be contained in the retroperitoneum for hours to weeks. These cases present a more difficult diagnostic problem but also a better chance of success if prompt action is taken [2,3]. However, these patients often present to internists who may not be familiar with the varied manifestations of a disease that is usually considered “surgical.” Despite the increased recognition of AA4 as a leading cause of death in the elderly population served by internists 141, the diagnosis of ruptured AA4 has received almost no attention in internal medicine journals in the last 20 years, and we are aware of no articles addressing how well internists perform at diagnosing ruptured AAA. To better identify the problems and pitfalls in the diagnosis of the subset of ruptured AA& initially seen by internists, we reviewed the recent experience at one medical center.
PATIENTS
AND METHODS
We reviewed operating room records, autopsy records, and computer records of discharge diagnoses at the Minneapolis Veterans Affairs (VA) Medical Center from January 1985 through June 1992. The charts of all patients identified as having ruptured AA4 were reviewed, and those whose outpatient or inpatient diagnostic evaluations were initially managed by internists were included in the series. The Minneapolis VA is a large acute-care academic medical center staffed primarily by resi1994
The American
Journal
of Medicine
Volume
96
163
RUPTURED
ABDOMINAL
AORTIC
ANEURYSM
/ LEDERLE
ET AL
TABLE I Clinical Features of 23 Patients With Ruptured Abdominal Aortic Aneurysm MAA)
TABLE II Incorrect Diagnosesin 23 Patients With Ruptured Abdominal Aortic Aneurysm MAAl*
% of Patients
Diagnosis
Symptoms Abdominal pain Back or flank pain Vomiting Transient improvement in symptoms Syncope Previously known AAA Groin pain Urinary retention Constipation Melena”
Urinary tract obstruction or infection Spinal disease Diverticulitis Colon cancer Appendicitis Constipation Renal stone Pulmonaryembolus ‘Some patients had more than one incorrect diagnosis
Signs Abdominal tenderness Pulsatile mass Systolic blood pressure < 110 mm Hg or orthostatic reduction 2 10 mm Hg Systolic blood pressure <90 mm Hg Abdominal distention Ecchymosis
ii :; 4
Laboratory findings Leukocyte count 2 1 1,000/mm3 Leukocyte count 2 20,000/mm3 Hemoglobin < 11 g/dL
:2” 4
rom aortoenteric flstula
dents. The medical center does not accept surgical emergencies and some patients are triaged directly to surgery in the emergency department, but most diagnostic evaluations are managed by internists. Thus the series was not intended to be representative of the entire spectrum of ruptured AAA but only of the generally less acute, diagnostically challenging subset seen by internists.
RESULTS We identified 23 patients with ruptured AAA whose diagnostic evaluations were managed by internists. Two patients were on the inpatient medicine service at the time of rupture. Three others were transferred to the inpatient medicine service for diagnostic evaluation after the onset of symptoms, 1 from the inpatient urology service and 2 from other hospitals. The remaining 18 patients were seen by internists in the emergency department. Ten of these were ultimately diagnosed in the emergency department and went to the operating room. The other eight remained undiagnosed in the emergency department and were admitted to medicine, in one case after the patient was discharged home and returned 4 days later. The patients were all men with a mean age of 72 years and a range of 59 to 91 years; only two were over 80 years old. Mean AAA size was 7.6 cm, with six aneurysms 5.5 cm or less at the time of rupture. Characteristics at presentation are shown in Table I. Twelve patients had symptoms of ruptured AAA 164
February
1994
The American
No. of Patients
Journal
of Medicine
Volume
for 24 hours or more before presenting to the Minneapolis VA, and 9 patients experienced an interval of 24 hours or more between presentation of AAA symptoms to the VA internist and surgery or death. Only three patients had a systolic blood pressure below 90 mm Hg at presentation to the internist. A pulsatile abdominal mass was palpated in 12 patients, but in only 4 did this finding precede radiologic diagnosis of AAA. Seventeen patients were taken to the operating room for emergency repair. Eight experienced shock (systolic pressure less than 60 mm Hg) prior to surgery, and 7 of these died, compared with 2 deaths in 9 patients without shock (p <0.02). All six patients who did not have surgery died, yielding an overall mortality of 65% for the series. In 14 cases the diagnosis of ruptured AAA was initially missed, resulting in delayed treatment while other diagnoses were investigated. The incorrect diagnoses entertained during these evaluations are shown in Table II. In eight cases, there was no indication from the chart that the diagnosis of symptomatic or ruptured AAA was suspected, whereas in the other six cases the diagnosis was considered but not felt to be the cause of the patient’s symptoms. In three of the latter, a computed tomography (CT) scan showing no leakage diverted attention from the AAA, that is, the diagnosis of symptomatic unruptured AAA was not made. In eight cases, the diagnosis of ruptured AAA was first suggested by a radiologic procedure done for reasons other than to evaluate AAA. The diagnosis was not made until after shock developed in 5 patients who were hemodynamically stable at presentation, and was made postmortem in another 5 patients, 4 of whom were hemodynamically stable at presentation. In four instances, cardiovascular collapse occurred while the patient was in the radiology suite. Although the evaluations in this series were primarily carried out by internists, it is perhaps indicative of the difficulty presented by these cases that consulting surgeons also failed to 96
RUPTURED
make a diagnosis of ruptured AAA in eight patients. Two of these consultations were at other hospitals prior to inpatient transfer and six occurred at the Minneapolis VA. Illustrative Cases CASE 1: A 63-year-old man presented to a VA satellite clinic with a history of 4 days of bilateral flank and lower abdominal pain, constipation, urinary frequency, urgency, dysuria, and nocturia. Blood pressure was 110/80 mm Hg, examination of the abdomen was normal but there was mild left flank tenderness. Leukocyte count was 12,900 per mm3, hemoglobin concentration was 11.7 g/dL, and the urine had 10 leukocytes per high-power field. The patient was admitted to the medicine service with a diagnosis of pyelonephritis. Abdominal radiography was nondiagnostic. The following morning, the patient was found pale, diaphoretic, and short of breath with a blood pressure of 75145 mm Hg and there was abdominal distension and tenderness. The patient noted that he had felt something “pop” in his abdomen. The abdomen was noted to be enlarging during resuscitation. Another abdominal radiograph was done to look for bowel obstruction and this revealed a calcified aorta and a possible AAA. Hemoglobin was now 3.5 g/dL. The patient was taken to the operating room in shock where ultrasound showed a lo-cm AAA. The aneurysm was repaired but the patient died of postoperative complications. CASE 2: A 71-year-old man was admitted to the hospital for treatment of an exacerbation of chronic obstructive lung disease. In the evening of the 25th hospital day, the patient developed severe left hip pain radiating to his left foot and was unable to urinate. There was tenderness over the lumbar spine. Radiographs of the spine showed no vertebral compression fractures. Left lower quadrant abdominal pain and tenderness developed on the 26th day and abdominal radiography showed a dilated colon and rectum. Leukocyte count was 17,900 per mm3, hemoglobin was 14.1 g/dL, and urinalysis was normal. The assessment was constipation, with possible diverticulitis. The patient was treated with laxatives and felt better the next day but his blood pressure was 100/58 mm Hg and he continued to have tenderness of the left lower quadrant and lumbosacral spine. An abdominal radiograph taken on the 28th hospital day showed a “possible aortic aneurysm” but this was considered an unlikely cause of the pain, which was still thought to be due to diverticulitis. Ultrasound was obtained nonurgently the following day and showed a 9.2-cm AAA. A CT scan confirmed bleeding into the left retroperitoneum. The patient developed February
ABDOMINAL
AORTIC
ANEURYSM
/ LEDERLE
ET AL
sudden severe worsening of his pain in the radiology suite and was taken directly to the operating room where the aneurysm was successfully repaired. CASE 3: A 77-year-old man came to the emergency department complaining of 4 days of back pain above the right hip radiating to the right abdomen and testicle with shortness of breath and one episode of vomiting. Blood pressure was 146190 mm Hg and there was guarding of the abdomen. Leukocyte count was 16,000 per mm3, hemoglobin was 11 g/dL, chest radiograph showed right lower lobe atelectasis, and abdominal radiograph showed calcification of the aorta and loss of definition of the right kidney and psoas muscle. The back and abdominal pain raised concern regarding AAA but the presentation was considered “atypical.” A consulting surgeon considered symptomatic AAA unlikely and the abdomen to be “not acutely surgical.” The patient was admitted to the medicine service with a diagnosis of back pain due to degenerative arthritis. A CT scan done nonurgently the next day showed a 7.6-cm AAA with hemorrhage into the right peritoneal space. The patient remained hemodynamically stable and was taken immediately to the operating room, where the AAA was successfully repaired.
COMMENTS We have reviewed the experience of patients with ruptured AAA presenting to internists at one medical center during a 7%year period. We may have missed some patients, for example, any who died of a ruptured AAA on a medical ward without diagnosis or autopsy, or who were sent out of the emergency department and either died at home or were readmitted to another hospital. Nevertheless, 23 cases, an average of 3 per year, were identified and most of these were initially misdiagnosed. These misdiagnoses primarily involved (1) not suspecting ruptured AAA as a cause of the patient’s symptoms, and (2) not giving the diagnosis high enough priority for prompt action in patients with a known AAA or in whom the diagnosis of ruptured AAA was considered during the course of the evaluation. Chart review revealed a general lack of physician awareness of the syndromes of contained rupture of AAA [2] and symptomatic unruptured AAA [31. With regard to the former, characteristic findings of contained rupture (urinary retention and flank pain, constipation, abdominal distension, leukocytosis, tenderness localized to an abdominal quadrant or radiating to the groin, absence of anemia or shock) were often considered to be evidence against the diagnosis. Regarding the symptomatic unrup1994
The American
Journal
of Medicine
Volume
96
165
RUPTURED XB~~~IINAL
AORTIC ANEURYSM / LEDERLE ET AL
tured MA, the association of pain with imminent rupture observed in this series is well known in the literature 131 and is attributed to acute expansion of the aneurysm or partial disruption of the pessel wall. Nevertheless, physicians frequently interpreted a CT scan showing AAA without leakage to mean that the AAA was an unlikely source of the patient’s symptoms. Another factor that contributed to delay in diagnosis was failure to palpate even a very large AA4 in an undistended abdomen until after it had been detected radiologically. We have previously found that routine abdominal palpation (not specifically directed at measuring the width of the aorta) routinely misses palpable pulsatile masses in asymptomatic patients [5]. The present series suggests that this is also the case with ruptured AAA. In addition, some ruptured AAAs are not palpable regardless of technique because of obesity, distension, guarding, or loss of integrity of the aneurysm. In our series, the fortuitous finding of evidence of AAA on plain radiographs ordered for other reasons was often the first clue to the diagnosis. Plain radiography, though not recommended for detecting asymptomatic AAA because of its dependence on calcification, has previously been shown (at least in retrospect) to provide diagnostic clues (calcification, soft-tissue mass, or loss of psoas or renal outlines) in up to 90% of cases of ruptured
surgical repair is obvious and is demonstrated by the marked increase in mortality associated with a delay of surgery until after shock ensues [I, 131, an association that was statistically significant even in our small series. Our study demonstrates the difficulty of applying the “classic triad” of abdominal pain, pulsatile mass, and hypotension to the diagnosis of ruptured AAA. The pulsatile mass is often missing or missed as noted above, hypotension often occurs as a late finding and marks the end of the best opportunity for a successful outcome, and abdominal pain has a long differential diagnosis which may lead the clinician away from urgent evaluation of the abdominal aorta. Our observations, in combination with previous reports, suggest recommendations for improving the diagnosis of ruptured AAA: (1) All physicians who provide care to older patients should become’ proficient at measuring the width of the aorta between the two index fingers. It is this width which provides evidence of MA, and not the prominence or force of the aortic pulsation, the presence of bruits, or other findings [5,71. (2) Physicians should lower their threshold for obtaining abdominal ultrasound in elderly patients with symptoms of abdominal or back pain. Screening asymptomatic elderly men for AAA with ultrasound has been widely (if not universally) recomAAA [61. mended [4,5], so it is difficult to justify failure to Recognition of the diagnostic difficulties associobtain ultrasound in a patient with symptoms ated with ruptured AAA is not new. Osler noted which could be caused by rupture or imminent that the diagnosis of retroperitoneal rupture was rupture of AAA. (3) A patient with abdominal pain particularly difficult and frequently overlooked 171. and shock should be taken directly to the operating He also noted that the symptoms could mimic room with ultrasound obtained while resuscitation renal or biliary colic and that constipation was a procedures are underway 110,131. (4) When rupcommon feature. Other authors have described tured AAA is suspected in a hemodynamically ruptured AAA presenting with urinary symptoms stable patient, an immediate CT scan is generally [8], gastrointestinal symptoms (abdominal pain considered to be safe and appropriate 110,131, and tenderness, constipation, distension, and, though it is of concern in this regard that four rarely, melena from aortoenteric fistula) [9,10], leukocytosis 1111, transient hypotension and syn- patients in our series suffered cardiovascular collapse in the radiology suite. (5) When a CT scan cope [ 12,131, ecchymoses of lower trunk or genitashows leakage from an AAA, the patient should be lia 181, and other more exotic presentations not taken immediately to the operating room. (6) The seen in our series such as inguinal hernia [141, finding of an AAA without leakage seen on CT scan femoral neuropathy [El, symptoms of cholecystishould not be interpreted as evidence that the tis 1161, and high output heart failure with hematuaneurysm is not the source of symptoms. The ria caused by rupture into the vena cava [ 171. possibility of symptomatic unruptured AAA must Although the various presentations of ruptured be given serious consideration even if the aneuAAA have been described in the literature, our rysm is small. These cases represent a difficult findings suggest that they may not be widely clinical decision in which symptoms, A4A size and appreciated by internists. If primary care physiother CT-scan findings, and the general health of cians are not familiar with the presentation of the patient must be assessed to determine whether ruptured AAA, it is unlikely that a vascular surto proceed promptly with surgery after adequate geon will be consulted in a timely manner. The importance of prompt diagnosis and immediate preoperative assessment aa preparation. 166
February 1994
The American Journal of Medicine
Volume 96
RUPTURED
In summary, we have found that although many patients with ruptured AAA present to internists, the diagnosis is often missed, largely due to a lack of awareness of the various manifestations of this condition. When shock develops during the resulting treatment delay, the chance of survival is markedly reduced. Internists should familiarize themselves with the presentation and management of ruptured AAA.
REFERENCES 1. Johansen K, Kohler TR, Nrcholls SC, Zierler RE, Clowes AW, Kazmers A. Ruptured abdominal aortic aneurysm: the Harborview experience. J Vast Surg 1991; 13: 240-7. 2. Jones CS, Reilly MK, Dalsing MC, Glover JL. Chronic contarned rupture of abdominal aorbc aneurysms. Arch Surg 1986; 121: 542-6. 3. Sullivan CA, Rohrer MJ, Cutler BS. Clinical management of the symptomatic but unruptured abdominal aortic aneurysm. J Vast Surg 1990; 11: 799-803. 4. Canadian Task Force on the Periodrc Health Examination. Periodic health examination, 1991 update: 5. Screening for abdominal aortic aneurysm. Can Med Assoc J 1991; 145: 783-9. 5. Lederle FA, WalkerJM, Reinke DB. Selectivescreenrngforabdominal aortrc
February
ABDOMINAL
AORTIC
ANEURYSM
/ LEDERLE
ET AL
aneurysms with physical examination and ultrasound. Arch Intern Med 1988; 148: 1753-6. 6. Loughran CF. A review of the plain abdominal radrograph in acute rupture of abdominal aortic aneurysms, Clin Radioi 1986; 37: 383-7. 7. Osler W. Aneurysm of the abdominal aorta. Lancet i905; 2: 1089-96. 8. Pryor JP. Diagnoses of ruptured aneurysm of abdominal aorta. BMJ 1972; 3: 735-6. 9. Sondhermer FK, Steinberg I. Gastrointestinal manifestations of abdomrnal aorbc aneurysms. AJR Am J Roentgen01 1964; 92: 1110-22. 10. Rutherford RB, McCroskey BL. Ruptured abdominal aortic aneurysms: special considerations. Surg Clin N Am 1989; 69: 859-68. 11. Mannick JA. Diagnosis of ruptured aneurysm of the abdominal aorta. N Engl J Med 1967; 276: 130557. 12. Moore HD. Diagnosis of rupture of abdominal aortic aneurysms. Lancet 1967; 2: 184-6. 13. Marston WA, Ahlquist R, Johnson G, Meyer AA. Misdiagnosis of ruptured abdominal aortic aneurysms. J VascSurg 1992; 16: 17-22. 14. Louras JC, Welch JP. Masking of ruptured abdominal aortic aneurysm by Incarcerated inguinal hernra. Arch Surg 1984; 119: 331-2. 15. Fletcher HS, Frankel J. Ruptured abdominal aneurysms presenting with unilateral peripheral neuropathy. Surgery 1976; 79: 120-l. 16. Chandler JJ. The Einstein sign: the clinical picture of acute cholecystitrs caused by ruptured abdomrnal aortic aneurysm [letter]. N Engl J Med 1984; 310: 1538. 17. Salo JA, Verkkala KA, Ala-Kulju KV, Heikkinen LO, Luosto RV. Hematuria is an indication of rupture of an abdominal aortic aneurysm into the vena cava. J Vast Surg 1990; 12: 41-4.
1994
The American
Journal
of Medicine
Volume
96
167
EDMUND
P.
PURPOSE: To define the clinical features and assess the frequency and causes of missed diagnoses of ruptured abdominal aortic aneurysm (AAA) in patients initially presenting to internists. PATIENTS: All identified patients with ruptured AAA presenting to internists during a 7V&year period at a large academic medical center. METHOD: Chart review. RESULTS: We identified 23 patients with a ruptured AAA presenting to internists. Most had abdominal pain and tenderness, back or flank pain, and leukocytosis, whereas anemia and profound hypotension (systolic blood pressure below 90 mm Hg) were uncommon at presentation. In 14 cases (61%), the diagnosis of ruptured AAA was initially missed. Nine patients had an interval of 24 hours or more between presentation to the internist and surgery or death. The diagnosis was not made until after shock developed in nine patients who were hemodynamically stable at presentation. Of 17 patients who underwent surgery, 7 of 8 with preoperative shock died, compared with 2 deaths in 9 patients (p c .02) without shock. All six patients who did not have surgery died, yielding an overall mortality of 65% for the series. Ruptured AAAs were most frequently misdiagnosed as urinary tract obstruction or infection, spinal disease, and diverticulitis. Chart review revealed a general lack of physician awareness of the syndromes of contained rupture of AAA and symptomatic unruptured AAA. CONCLUSIONS: In patients with ruptured AAA who present to internists, the diagno-
From the Department of Medicine (FAL, CMP), the Department of Surgery (EPC), Minneapolis Veterans Affairs Medical Center, University of Minnesota, Minneapolis, Minnesota. Requests for reprints should be addressed to Frank A. Lederle, M.D., Minneapolis Veterans Affairs Medical Center, Department of Medicine (III-O), OneVeterans Drive, Minneapolis, Minnesota 55417. Manuscript submitted December 29, 1992, and accepted in revised form March 31, 1993.
February
CHUTE,
M.D.,
Minneapo/is,
Minnesota
sis is often delayed or missed and this appears to adversely effect survival. Internists should familiarize themselves with the presentation and management of ruptured AAA.
T
here are few conditions encountered in the practice of medicine that require accurate diagnosis more urgently than does rupture of an abdominal aortic aneurysm L4AA). In many patients, ruptured AA4 presents as sudden vascular collapse, in which case the need for emergency surgery may be obvious but the likelihood of a successful outcome is low [ll. In other patients, an AAA may become symptomatic prior to rupture or an aneurysmal leak may be contained in the retroperitoneum for hours to weeks. These cases present a more difficult diagnostic problem but also a better chance of success if prompt action is taken [2,3]. However, these patients often present to internists who may not be familiar with the varied manifestations of a disease that is usually considered “surgical.” Despite the increased recognition of AA4 as a leading cause of death in the elderly population served by internists 141, the diagnosis of ruptured AA4 has received almost no attention in internal medicine journals in the last 20 years, and we are aware of no articles addressing how well internists perform at diagnosing ruptured AAA. To better identify the problems and pitfalls in the diagnosis of the subset of ruptured AA& initially seen by internists, we reviewed the recent experience at one medical center.
PATIENTS
AND METHODS
We reviewed operating room records, autopsy records, and computer records of discharge diagnoses at the Minneapolis Veterans Affairs (VA) Medical Center from January 1985 through June 1992. The charts of all patients identified as having ruptured AA4 were reviewed, and those whose outpatient or inpatient diagnostic evaluations were initially managed by internists were included in the series. The Minneapolis VA is a large acute-care academic medical center staffed primarily by resi1994
The American
Journal
of Medicine
Volume
96
163
RUPTURED
ABDOMINAL
AORTIC
ANEURYSM
/ LEDERLE
ET AL
TABLE I Clinical Features of 23 Patients With Ruptured Abdominal Aortic Aneurysm MAA)
TABLE II Incorrect Diagnosesin 23 Patients With Ruptured Abdominal Aortic Aneurysm MAAl*
% of Patients
Diagnosis
Symptoms Abdominal pain Back or flank pain Vomiting Transient improvement in symptoms Syncope Previously known AAA Groin pain Urinary retention Constipation Melena”
Urinary tract obstruction or infection Spinal disease Diverticulitis Colon cancer Appendicitis Constipation Renal stone Pulmonaryembolus ‘Some patients had more than one incorrect diagnosis
Signs Abdominal tenderness Pulsatile mass Systolic blood pressure < 110 mm Hg or orthostatic reduction 2 10 mm Hg Systolic blood pressure <90 mm Hg Abdominal distention Ecchymosis
ii :; 4
Laboratory findings Leukocyte count 2 1 1,000/mm3 Leukocyte count 2 20,000/mm3 Hemoglobin < 11 g/dL
:2” 4
rom aortoenteric flstula
dents. The medical center does not accept surgical emergencies and some patients are triaged directly to surgery in the emergency department, but most diagnostic evaluations are managed by internists. Thus the series was not intended to be representative of the entire spectrum of ruptured AAA but only of the generally less acute, diagnostically challenging subset seen by internists.
RESULTS We identified 23 patients with ruptured AAA whose diagnostic evaluations were managed by internists. Two patients were on the inpatient medicine service at the time of rupture. Three others were transferred to the inpatient medicine service for diagnostic evaluation after the onset of symptoms, 1 from the inpatient urology service and 2 from other hospitals. The remaining 18 patients were seen by internists in the emergency department. Ten of these were ultimately diagnosed in the emergency department and went to the operating room. The other eight remained undiagnosed in the emergency department and were admitted to medicine, in one case after the patient was discharged home and returned 4 days later. The patients were all men with a mean age of 72 years and a range of 59 to 91 years; only two were over 80 years old. Mean AAA size was 7.6 cm, with six aneurysms 5.5 cm or less at the time of rupture. Characteristics at presentation are shown in Table I. Twelve patients had symptoms of ruptured AAA 164
February
1994
The American
No. of Patients
Journal
of Medicine
Volume
for 24 hours or more before presenting to the Minneapolis VA, and 9 patients experienced an interval of 24 hours or more between presentation of AAA symptoms to the VA internist and surgery or death. Only three patients had a systolic blood pressure below 90 mm Hg at presentation to the internist. A pulsatile abdominal mass was palpated in 12 patients, but in only 4 did this finding precede radiologic diagnosis of AAA. Seventeen patients were taken to the operating room for emergency repair. Eight experienced shock (systolic pressure less than 60 mm Hg) prior to surgery, and 7 of these died, compared with 2 deaths in 9 patients without shock (p <0.02). All six patients who did not have surgery died, yielding an overall mortality of 65% for the series. In 14 cases the diagnosis of ruptured AAA was initially missed, resulting in delayed treatment while other diagnoses were investigated. The incorrect diagnoses entertained during these evaluations are shown in Table II. In eight cases, there was no indication from the chart that the diagnosis of symptomatic or ruptured AAA was suspected, whereas in the other six cases the diagnosis was considered but not felt to be the cause of the patient’s symptoms. In three of the latter, a computed tomography (CT) scan showing no leakage diverted attention from the AAA, that is, the diagnosis of symptomatic unruptured AAA was not made. In eight cases, the diagnosis of ruptured AAA was first suggested by a radiologic procedure done for reasons other than to evaluate AAA. The diagnosis was not made until after shock developed in 5 patients who were hemodynamically stable at presentation, and was made postmortem in another 5 patients, 4 of whom were hemodynamically stable at presentation. In four instances, cardiovascular collapse occurred while the patient was in the radiology suite. Although the evaluations in this series were primarily carried out by internists, it is perhaps indicative of the difficulty presented by these cases that consulting surgeons also failed to 96
RUPTURED
make a diagnosis of ruptured AAA in eight patients. Two of these consultations were at other hospitals prior to inpatient transfer and six occurred at the Minneapolis VA. Illustrative Cases CASE 1: A 63-year-old man presented to a VA satellite clinic with a history of 4 days of bilateral flank and lower abdominal pain, constipation, urinary frequency, urgency, dysuria, and nocturia. Blood pressure was 110/80 mm Hg, examination of the abdomen was normal but there was mild left flank tenderness. Leukocyte count was 12,900 per mm3, hemoglobin concentration was 11.7 g/dL, and the urine had 10 leukocytes per high-power field. The patient was admitted to the medicine service with a diagnosis of pyelonephritis. Abdominal radiography was nondiagnostic. The following morning, the patient was found pale, diaphoretic, and short of breath with a blood pressure of 75145 mm Hg and there was abdominal distension and tenderness. The patient noted that he had felt something “pop” in his abdomen. The abdomen was noted to be enlarging during resuscitation. Another abdominal radiograph was done to look for bowel obstruction and this revealed a calcified aorta and a possible AAA. Hemoglobin was now 3.5 g/dL. The patient was taken to the operating room in shock where ultrasound showed a lo-cm AAA. The aneurysm was repaired but the patient died of postoperative complications. CASE 2: A 71-year-old man was admitted to the hospital for treatment of an exacerbation of chronic obstructive lung disease. In the evening of the 25th hospital day, the patient developed severe left hip pain radiating to his left foot and was unable to urinate. There was tenderness over the lumbar spine. Radiographs of the spine showed no vertebral compression fractures. Left lower quadrant abdominal pain and tenderness developed on the 26th day and abdominal radiography showed a dilated colon and rectum. Leukocyte count was 17,900 per mm3, hemoglobin was 14.1 g/dL, and urinalysis was normal. The assessment was constipation, with possible diverticulitis. The patient was treated with laxatives and felt better the next day but his blood pressure was 100/58 mm Hg and he continued to have tenderness of the left lower quadrant and lumbosacral spine. An abdominal radiograph taken on the 28th hospital day showed a “possible aortic aneurysm” but this was considered an unlikely cause of the pain, which was still thought to be due to diverticulitis. Ultrasound was obtained nonurgently the following day and showed a 9.2-cm AAA. A CT scan confirmed bleeding into the left retroperitoneum. The patient developed February
ABDOMINAL
AORTIC
ANEURYSM
/ LEDERLE
ET AL
sudden severe worsening of his pain in the radiology suite and was taken directly to the operating room where the aneurysm was successfully repaired. CASE 3: A 77-year-old man came to the emergency department complaining of 4 days of back pain above the right hip radiating to the right abdomen and testicle with shortness of breath and one episode of vomiting. Blood pressure was 146190 mm Hg and there was guarding of the abdomen. Leukocyte count was 16,000 per mm3, hemoglobin was 11 g/dL, chest radiograph showed right lower lobe atelectasis, and abdominal radiograph showed calcification of the aorta and loss of definition of the right kidney and psoas muscle. The back and abdominal pain raised concern regarding AAA but the presentation was considered “atypical.” A consulting surgeon considered symptomatic AAA unlikely and the abdomen to be “not acutely surgical.” The patient was admitted to the medicine service with a diagnosis of back pain due to degenerative arthritis. A CT scan done nonurgently the next day showed a 7.6-cm AAA with hemorrhage into the right peritoneal space. The patient remained hemodynamically stable and was taken immediately to the operating room, where the AAA was successfully repaired.
COMMENTS We have reviewed the experience of patients with ruptured AAA presenting to internists at one medical center during a 7%year period. We may have missed some patients, for example, any who died of a ruptured AAA on a medical ward without diagnosis or autopsy, or who were sent out of the emergency department and either died at home or were readmitted to another hospital. Nevertheless, 23 cases, an average of 3 per year, were identified and most of these were initially misdiagnosed. These misdiagnoses primarily involved (1) not suspecting ruptured AAA as a cause of the patient’s symptoms, and (2) not giving the diagnosis high enough priority for prompt action in patients with a known AAA or in whom the diagnosis of ruptured AAA was considered during the course of the evaluation. Chart review revealed a general lack of physician awareness of the syndromes of contained rupture of AAA [2] and symptomatic unruptured AAA [31. With regard to the former, characteristic findings of contained rupture (urinary retention and flank pain, constipation, abdominal distension, leukocytosis, tenderness localized to an abdominal quadrant or radiating to the groin, absence of anemia or shock) were often considered to be evidence against the diagnosis. Regarding the symptomatic unrup1994
The American
Journal
of Medicine
Volume
96
165
RUPTURED XB~~~IINAL
AORTIC ANEURYSM / LEDERLE ET AL
tured MA, the association of pain with imminent rupture observed in this series is well known in the literature 131 and is attributed to acute expansion of the aneurysm or partial disruption of the pessel wall. Nevertheless, physicians frequently interpreted a CT scan showing AAA without leakage to mean that the AAA was an unlikely source of the patient’s symptoms. Another factor that contributed to delay in diagnosis was failure to palpate even a very large AA4 in an undistended abdomen until after it had been detected radiologically. We have previously found that routine abdominal palpation (not specifically directed at measuring the width of the aorta) routinely misses palpable pulsatile masses in asymptomatic patients [5]. The present series suggests that this is also the case with ruptured AAA. In addition, some ruptured AAAs are not palpable regardless of technique because of obesity, distension, guarding, or loss of integrity of the aneurysm. In our series, the fortuitous finding of evidence of AAA on plain radiographs ordered for other reasons was often the first clue to the diagnosis. Plain radiography, though not recommended for detecting asymptomatic AAA because of its dependence on calcification, has previously been shown (at least in retrospect) to provide diagnostic clues (calcification, soft-tissue mass, or loss of psoas or renal outlines) in up to 90% of cases of ruptured
surgical repair is obvious and is demonstrated by the marked increase in mortality associated with a delay of surgery until after shock ensues [I, 131, an association that was statistically significant even in our small series. Our study demonstrates the difficulty of applying the “classic triad” of abdominal pain, pulsatile mass, and hypotension to the diagnosis of ruptured AAA. The pulsatile mass is often missing or missed as noted above, hypotension often occurs as a late finding and marks the end of the best opportunity for a successful outcome, and abdominal pain has a long differential diagnosis which may lead the clinician away from urgent evaluation of the abdominal aorta. Our observations, in combination with previous reports, suggest recommendations for improving the diagnosis of ruptured AAA: (1) All physicians who provide care to older patients should become’ proficient at measuring the width of the aorta between the two index fingers. It is this width which provides evidence of MA, and not the prominence or force of the aortic pulsation, the presence of bruits, or other findings [5,71. (2) Physicians should lower their threshold for obtaining abdominal ultrasound in elderly patients with symptoms of abdominal or back pain. Screening asymptomatic elderly men for AAA with ultrasound has been widely (if not universally) recomAAA [61. mended [4,5], so it is difficult to justify failure to Recognition of the diagnostic difficulties associobtain ultrasound in a patient with symptoms ated with ruptured AAA is not new. Osler noted which could be caused by rupture or imminent that the diagnosis of retroperitoneal rupture was rupture of AAA. (3) A patient with abdominal pain particularly difficult and frequently overlooked 171. and shock should be taken directly to the operating He also noted that the symptoms could mimic room with ultrasound obtained while resuscitation renal or biliary colic and that constipation was a procedures are underway 110,131. (4) When rupcommon feature. Other authors have described tured AAA is suspected in a hemodynamically ruptured AAA presenting with urinary symptoms stable patient, an immediate CT scan is generally [8], gastrointestinal symptoms (abdominal pain considered to be safe and appropriate 110,131, and tenderness, constipation, distension, and, though it is of concern in this regard that four rarely, melena from aortoenteric fistula) [9,10], leukocytosis 1111, transient hypotension and syn- patients in our series suffered cardiovascular collapse in the radiology suite. (5) When a CT scan cope [ 12,131, ecchymoses of lower trunk or genitashows leakage from an AAA, the patient should be lia 181, and other more exotic presentations not taken immediately to the operating room. (6) The seen in our series such as inguinal hernia [141, finding of an AAA without leakage seen on CT scan femoral neuropathy [El, symptoms of cholecystishould not be interpreted as evidence that the tis 1161, and high output heart failure with hematuaneurysm is not the source of symptoms. The ria caused by rupture into the vena cava [ 171. possibility of symptomatic unruptured AAA must Although the various presentations of ruptured be given serious consideration even if the aneuAAA have been described in the literature, our rysm is small. These cases represent a difficult findings suggest that they may not be widely clinical decision in which symptoms, A4A size and appreciated by internists. If primary care physiother CT-scan findings, and the general health of cians are not familiar with the presentation of the patient must be assessed to determine whether ruptured AAA, it is unlikely that a vascular surto proceed promptly with surgery after adequate geon will be consulted in a timely manner. The importance of prompt diagnosis and immediate preoperative assessment aa preparation. 166
February 1994
The American Journal of Medicine
Volume 96
RUPTURED
In summary, we have found that although many patients with ruptured AAA present to internists, the diagnosis is often missed, largely due to a lack of awareness of the various manifestations of this condition. When shock develops during the resulting treatment delay, the chance of survival is markedly reduced. Internists should familiarize themselves with the presentation and management of ruptured AAA.
REFERENCES 1. Johansen K, Kohler TR, Nrcholls SC, Zierler RE, Clowes AW, Kazmers A. Ruptured abdominal aortic aneurysm: the Harborview experience. J Vast Surg 1991; 13: 240-7. 2. Jones CS, Reilly MK, Dalsing MC, Glover JL. Chronic contarned rupture of abdominal aorbc aneurysms. Arch Surg 1986; 121: 542-6. 3. Sullivan CA, Rohrer MJ, Cutler BS. Clinical management of the symptomatic but unruptured abdominal aortic aneurysm. J Vast Surg 1990; 11: 799-803. 4. Canadian Task Force on the Periodrc Health Examination. Periodic health examination, 1991 update: 5. Screening for abdominal aortic aneurysm. Can Med Assoc J 1991; 145: 783-9. 5. Lederle FA, WalkerJM, Reinke DB. Selectivescreenrngforabdominal aortrc
February
ABDOMINAL
AORTIC
ANEURYSM
/ LEDERLE
ET AL
aneurysms with physical examination and ultrasound. Arch Intern Med 1988; 148: 1753-6. 6. Loughran CF. A review of the plain abdominal radrograph in acute rupture of abdominal aortic aneurysms, Clin Radioi 1986; 37: 383-7. 7. Osler W. Aneurysm of the abdominal aorta. Lancet i905; 2: 1089-96. 8. Pryor JP. Diagnoses of ruptured aneurysm of abdominal aorta. BMJ 1972; 3: 735-6. 9. Sondhermer FK, Steinberg I. Gastrointestinal manifestations of abdomrnal aorbc aneurysms. AJR Am J Roentgen01 1964; 92: 1110-22. 10. Rutherford RB, McCroskey BL. Ruptured abdominal aortic aneurysms: special considerations. Surg Clin N Am 1989; 69: 859-68. 11. Mannick JA. Diagnosis of ruptured aneurysm of the abdominal aorta. N Engl J Med 1967; 276: 130557. 12. Moore HD. Diagnosis of rupture of abdominal aortic aneurysms. Lancet 1967; 2: 184-6. 13. Marston WA, Ahlquist R, Johnson G, Meyer AA. Misdiagnosis of ruptured abdominal aortic aneurysms. J VascSurg 1992; 16: 17-22. 14. Louras JC, Welch JP. Masking of ruptured abdominal aortic aneurysm by Incarcerated inguinal hernra. Arch Surg 1984; 119: 331-2. 15. Fletcher HS, Frankel J. Ruptured abdominal aneurysms presenting with unilateral peripheral neuropathy. Surgery 1976; 79: 120-l. 16. Chandler JJ. The Einstein sign: the clinical picture of acute cholecystitrs caused by ruptured abdomrnal aortic aneurysm [letter]. N Engl J Med 1984; 310: 1538. 17. Salo JA, Verkkala KA, Ala-Kulju KV, Heikkinen LO, Luosto RV. Hematuria is an indication of rupture of an abdominal aortic aneurysm into the vena cava. J Vast Surg 1990; 12: 41-4.
1994
The American
Journal
of Medicine
Volume
96
167